Ultimately, I think we want to allow alpha numeric characters only, replacing certain special characters including spaces with hyphens, and stripping out others.

Wordpress does this with a function called sanitize_title_with_dashes(), the source of which can be found here - https://core.trac.wordpress.org/browser/tags/4.0.1/src/wp-includes/formatting.php#L0 Reviewing this function may be useful in determining the best approach.

I looked at WP when adding in the KB. They use remove_accents() to convert UTF8 characters to ASCII, then only use those filtered alphanumeric ASCII characters in the title.

Since we're allowing UTF8 characters, each of those will be broken into its corresponding octet (e.g. %21 is an exclamation point), and the browser determines how to display the correct characters by decoding each octet or not. If we were to remove the possibility of encoded characters in URLs, then WP's solution would be better since it only allows ASCII. However, removing each of the above characters is merely our happy medium.

This has been updated to remove non-alphanumeric characters from the ASCII table, #0-127. Also removes consecutive dashes that may result from consecutively removed characters separated by spaces.

Going to mark this resolved, unless there is more you'd like to update here?

That sounds fine with me